Farm-level climate adaptation and socio-economic determinants: A CART analysis of farmers in Haryana, India

Climate change poses serious challenges to agricultural sustainability in India's semi-arid regions, particularly in intensively cultivated states such as Haryana. Despite growing interest in farm-level climate adaptation, empirical evidence on the non-linear, interactive socioeconomic determinants of adoption decisions in semi-arid South Asia remains limited. This study addresses this gap by analysing crop production-related climate adaptation strategies and their socio-economic determinants among 120 farmers in Karnal district of Haryana, India, using Classification and Regression Tree (CART) analysis and K-means cluster analysis. Farmers were selected through a multistage sampling design comprising equal numbers of beneficiaries and non-beneficiaries receiving climate services under National Innovations in Climate Resilient Agriculture (NICRA) project. Primary data were collected on socioeconomic characteristics, institutional access, and adoption of sixteen climate-resilient practices. Adoption was higher among NICRA beneficiaries for fifteen of sixteen practices, integrated nutrient management being the sole exception. CART identified farming experience, family size, mass media exposure, family education status, age and operational landholding as principal splitting variables. Farming experience most often formed the root node, governing knowledge-intensive practices, whereas family size primarily determined high-yielding varieties, timely sowing, irrigation and zero tillage, with extension contact entering secondarily. Mass media exposure led water management and crop rotation, and landholding contingency cropping alone. Cluster analysis revealed three different farmer typologies- high-adaptation responsive, strategy-specific, and low-adaptation conventional differentiated by the intensity and combination of strategies adopted, highlighting pronounced heterogeneity in adaptive behaviour. These findings provide an empirical basis for designing targeted, typology-specific climate service delivery and extension interventions.